Erase to Enhance: Data-Efficient Machine Unlearning in MRI Reconstruction

Yuyang Xue, Jingshuai Liu, Steven McDonagh, Sotirios A. Tsaftaris
Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, PMLR 250:1785-1800, 2024.

Abstract

Machine unlearning is a promising paradigm for removing unwanted data samples from a trained model, towards ensuring compliance with privacy regulations and limiting harmful biases. Although unlearning has been shown in, e.g., classification and recommendation systems, its potential in medical image-to-image translation, specifically in image reconstruction, has not been thoroughly investigated. This paper shows that machine unlearning is possible in MRI tasks and has the potential to benefit for bias removal. We set up a protocol to study how much shared knowledge exists between datasets of different organs, allowing us to effectively quantify the effect of unlearning. Our study reveals that combining training data can lead to hallucinations and reduced image quality in the reconstructed data. We use unlearning to remove hallucinations as a proxy exemplar of undesired data removal. Indeed, we show that machine unlearning is possible without full retraining. Furthermore, our observations indicate that maintaining high performance is feasible even when using only a subset of retain data. We have made our code publicly accessible.

Cite this Paper

BibTeX
@InProceedings{pmlr-v250-xue24a, title = {Erase to Enhance: Data-Efficient Machine Unlearning in MRI Reconstruction}, author = {Xue, Yuyang and Liu, Jingshuai and McDonagh, Steven and Tsaftaris, Sotirios A.}, booktitle = {Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning}, pages = {1785--1800}, year = {2024}, editor = {Burgos, Ninon and Petitjean, Caroline and Vakalopoulou, Maria and Christodoulidis, Stergios and Coupe, Pierrick and Delingette, Hervé and Lartizien, Carole and Mateus, Diana}, volume = {250}, series = {Proceedings of Machine Learning Research}, month = {03--05 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v250/main/assets/xue24a/xue24a.pdf}, url = {https://proceedings.mlr.press/v250/xue24a.html}, abstract = {Machine unlearning is a promising paradigm for removing unwanted data samples from a trained model, towards ensuring compliance with privacy regulations and limiting harmful biases. Although unlearning has been shown in, e.g., classification and recommendation systems, its potential in medical image-to-image translation, specifically in image reconstruction, has not been thoroughly investigated. This paper shows that machine unlearning is possible in MRI tasks and has the potential to benefit for bias removal. We set up a protocol to study how much shared knowledge exists between datasets of different organs, allowing us to effectively quantify the effect of unlearning. Our study reveals that combining training data can lead to hallucinations and reduced image quality in the reconstructed data. We use unlearning to remove hallucinations as a proxy exemplar of undesired data removal. Indeed, we show that machine unlearning is possible without full retraining. Furthermore, our observations indicate that maintaining high performance is feasible even when using only a subset of retain data. We have made our code publicly accessible.} }
Endnote
%0 Conference Paper %T Erase to Enhance: Data-Efficient Machine Unlearning in MRI Reconstruction %A Yuyang Xue %A Jingshuai Liu %A Steven McDonagh %A Sotirios A. Tsaftaris %B Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning %C Proceedings of Machine Learning Research %D 2024 %E Ninon Burgos %E Caroline Petitjean %E Maria Vakalopoulou %E Stergios Christodoulidis %E Pierrick Coupe %E Hervé Delingette %E Carole Lartizien %E Diana Mateus %F pmlr-v250-xue24a %I PMLR %P 1785--1800 %U https://proceedings.mlr.press/v250/xue24a.html %V 250 %X Machine unlearning is a promising paradigm for removing unwanted data samples from a trained model, towards ensuring compliance with privacy regulations and limiting harmful biases. Although unlearning has been shown in, e.g., classification and recommendation systems, its potential in medical image-to-image translation, specifically in image reconstruction, has not been thoroughly investigated. This paper shows that machine unlearning is possible in MRI tasks and has the potential to benefit for bias removal. We set up a protocol to study how much shared knowledge exists between datasets of different organs, allowing us to effectively quantify the effect of unlearning. Our study reveals that combining training data can lead to hallucinations and reduced image quality in the reconstructed data. We use unlearning to remove hallucinations as a proxy exemplar of undesired data removal. Indeed, we show that machine unlearning is possible without full retraining. Furthermore, our observations indicate that maintaining high performance is feasible even when using only a subset of retain data. We have made our code publicly accessible.
APA
Xue, Y., Liu, J., McDonagh, S. & Tsaftaris, S.A.. (2024). Erase to Enhance: Data-Efficient Machine Unlearning in MRI Reconstruction. Proceedings of The 7nd International Conference on Medical Imaging with Deep Learning, in Proceedings of Machine Learning Research 250:1785-1800 Available from https://proceedings.mlr.press/v250/xue24a.html.

Related Material